The same problem still exists in the Openseespymac version. Is it possible to solve the issue in macs?

[serkanhasanoglu]

The same problem still exists in the Openseespymac version. Is it possible to solve the issue in macs?

Originally posted by @serkanhasanoglu in #121 (comment)

[zhuminjie]

Is it a mac issue or OpenSeesPy issue?

[volkanozsarac]

I guess Mac version of OpenSeesPy has not been compiled recently. The bug has been fixed only recently in the latest version. In Linux and Windows OS, while using latest version of OpenSeesPy, the command is working fine.

[zhuminjie]

I see. I am planning to update the mac version for arm64 processors, but not sure how many users are still using intel

[volkanozsarac]

Since people started finding other workarounds such as use of Rosetta terminal, I think the update would still be super useful!

[serkanhasanoglu]

Openseespy currently doesn't work with mac version for arm64 processors. Hence, as i mac user, i use Rosetta terminal which provides the ability to use intel-based applications on new Macs with arm64 processors. It would be great if you update the Mac version for arm64 processors.

[ihsanenginbal]

This can be solved with a workaround because Miniconda also creates x86 environments.

• Install miniconda
• Create an environment configured for x86, follow the below steps in the terminal:

conda create -n my_x86_env -y
conda activate my_x86_env
conda config --env --set subdir osx-64

• Install the Python version (stay compatible with Spyder if you will continue with Spyder later, so at the time of writing this entry, do not go above Python v11!)

conda install python=3.10

• Install packages using conda or pip

conda install numpy
pip install networkx

pip install also the openseespy for mac

Once this environment is active, whatever you pip install will always be pulled from x86 architecture.

The rest is easier for Anaconda users. Once this environment is created, it will automatically appear in the "Environments" list of the Anaconda. Click that environment, then go to Home of Anaconda and install Spyder over that environment. The rest is business as usual.

I think the same can be done also in PyCharm by using this new environment, but I have not tried that one.

x86 version of course runs slower, by the way.

P.S.: A part of my solution was taken from here:
https://stackoverflow.com/questions/71691598/how-to-run-python-as-x86-with-rosetta2-on-arm-macos-machine

[serkanhasanoglu]

@ihsanenginbal thanks for the information. Using a conda environment finds solutions to using openseespy in macs. The problem here actually is that the latest "openseespymac" version has not been updated since Jul 22, 2022. Hence, new changes in openseespy commands that were implemented after this date, cannot be used and cause problems.

[cschthy]

Is it a mac issue or OpenSeesPy issue?

I see. I am planning to update the mac version for arm64 processors, but not sure how many users are still using intel

Hello,the openseesmac on M1 has been solved?

[edddieeeee]

@ihsanenginbal thanks for the information. Using a conda environment finds solutions to using openseespy in macs. The problem here actually is that the latest "openseespymac" version has not been updated since Jul 22, 2022. Hence, new changes in openseespy commands that were implemented after this date, cannot be used and cause problems.

hi i saw your solution before but i can not apply can you explain more with photo

[odarroyo]

I see. I am planning to update the mac version for arm64 processors, but not sure how many users are still using intel

Hi! I definitively see it as a worthy endeavor. The base use for OpenSeesPy using Intel is probably small by now, given that the M1 processor was launched in 2020. Moreover, even through Rosetta, OpenSeesPy runs pretty fast on the M1/M2/M3 Macs that I think they deserve their native version.

[odarroyo]

This can be solved with a workaround because Miniconda also creates x86 environments.

• Install miniconda
• Create an environment configured for x86, follow the below steps in the terminal:

conda create -n my_x86_env -y conda activate my_x86_env conda config --env --set subdir osx-64

• Install the Python version (stay compatible with Spyder if you will continue with Spyder later, so at the time of writing this entry, do not go above Python v11!)

conda install python=3.10

• Install packages using conda or pip

conda install numpy pip install networkx

pip install also the openseespy for mac

Once this environment is active, whatever you pip install will always be pulled from x86 architecture.

The rest is easier for Anaconda users. Once this environment is created, it will automatically appear in the "Environments" list of the Anaconda. Click that environment, then go to Home of Anaconda and install Spyder over that environment. The rest is business as usual.

I think the same can be done also in PyCharm by using this new environment, but I have not tried that one.

x86 version of course runs slower, by the way.

P.S.: A part of my solution was taken from here: https://stackoverflow.com/questions/71691598/how-to-run-python-as-x86-with-rosetta2-on-arm-macos-machine

Thank you for posting that solution, it was really useful.

[serkanhasanoglu]

I see. I am planning to update the mac version for arm64 processors, but not sure how many users are still using intel

@zhuminjie Is there any progress on this?

[odarroyo]

Same question as @serkanhasanoglu. I've performed some testing running pushover analyses of 3D frames and the Arm versions are quite fast. An M3 Pro processor runs as fast as a Ryzen 9 7950X even with the emulation overload, so it is possible that it may beat it if run natively.